AI Tools: A Dictionary
Artificial Intelligence (AI): a machine’s ability to perform a task that might previously have required human intelligence and intervention
Black box: machine learning models that produce a result or response without explaining or showing how they did so. The internal processes used and the various weighted factors remain unknown. Technology with a lack of transparency.
Code: instructions for software
Generative AI: tools that generate words, images or sounds in response to prompts with human-like efficiency.
Hallucination: when a large language model generates false information. An LLM “hallucinating” can mean it is generating inaccurate historical facts, creating a nonsensical answer, or putting words or images together that seem to make sense but do not.
Inputs: also known as prompts. The signals or data received by an AI tool
Outputs: the signals or data sent from an AI tool.
Large Language Model (LLM): a deep learning algorithm that can recognize, summarize, translate, predict and generate text and other forms of content based on information contained in massive datasets. These datasets include whole novels, research papers, webpages, advertisements, Reddit posts, and much more. ChatGPT and Google Bard are well known examples of LLMs.
Machine learning: a component of creating artificial intelligence (AI) that focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its predictive accuracy.
Multimodal AI: an advanced form of generative AI tools that analyzes and interprets multiple modes of data simultaneously, allowing it to generate more accurate and conversational responses. Multimodal AI can combine information from different source types, such as images, sound, and text, to create a more thorough response to a given prompt.
Neural network: a set of algorithms, modeled loosely after the human brain, that are designed to recognize patterns.
Prompt: any form of text, question, information, or coding that communicates to AI what response you’re looking for. We also like the definition of prompts a recent Parsons graduate gave us: “Prompts are code in natural language.”
Prompt engineering: the practice of crafting effective queries or inputs — referred to as prompts — to guide an AI tool towards generating desired responses.
Transformer architecture model: a neural network that is able to identify context and thus meaning by tracking relationships in sequential data like the words in this sentence. This is what enables a generative AI tool to simultaneously analyze quantities of data instead of going word by word.